Artificial Intelligence News & Discussion

Was asked by a family member to check out this Netflix sponsored AI robotics film. Some featured within are well into the soup of the military industrial complex, no surprise. Ukraine is mentioned a number of times so it is relatively current (2 months it says), with a Russia bad, we must be good aim.

One section looked to the question of AI producing medical drugs to help save lives, which is reasonable, yet here is the thing, the programming question had a 1 at the end (the 0 or 1), and when they changed it to a 0 and left it for AI computation overnight, the next morning there were 40,000 extremely toxic chemical formulas produced. It was said, anyone can produce this result - full chemical equations.

Here is the trailer, with the Netflix version around 1 hour in length. It left me feeling a wee bit hollow to say the least. There seemed to be a priming aspect to it.

 
Very interesting talk on many things but AI is talked about for the second half of the podcast.

Marc Andreessen from WIKI:​

is an American entrepreneur, venture capital investor, and software engineer. He is the co-author of Mosaic, the first widely used web browser with a graphical user interface; co-founder of Netscape;[2] and co-founder and general partner of Silicon Valley venture capital firm Andreessen Horowitz. He co-founded and later sold the software company Opsware to Hewlett-Packard. Andreessen is also a co-founder of Ning, a company that provides a platform for social networking websites. He sits on the board of directors of Meta Platforms. Andreessen was one of six inductees in the World Wide Web Hall of Fame announced at the First International Conference on the World-Wide Web in 1994.[3][4] As of February 2023, his net-worth is estimated at $1.7 billion by Forbes.[5] In March 2022 he was appointed to the Homeland Security Advisory Council by Secretary of Homeland Security Alejandro Mayorkas[6]


 
RT is banned from France, I can't access the article. May you copy the content here ?

Russian AI bot outperforms ChatGPT – developers​

9 Sep, 2023
YandexGPT “steadily surpasses” its US competitor, its creators have claimed

A new Russian AI bot has shown greater promise than the most popular chatbot developed by US-based OpenAI, IT giant Yandex claimed in comments to RIA Novosti on Saturday.

Dubbed YandexGPT, the Russian bot’s “basic model steadily surpasses” the US ChatGPT 3.5 version when it comes to generating answers in Russian, the company’s search and advertisement technologies business group director, Dmitry Masyuk, said.

The Russian bot also “provides a higher quality answer than ChatGPT 4” in many cases, Masyuk said, adding that it is “only a matter of time” before YandexGPT could rival its US competitor. When compared with LLama 2 7b, another US chatbot developed by Facebook parent company Meta, the Russian bot already provides better results, even in generating answers in English, he claimed.

Yandex constantly evaluates the progress of its AI product to ChatGPT, Masyuk said, adding that it is hard to compare the two systems since they can be better at different things. “If a neural network that solves tasks in physics, writes a fairy tale and can make a letter to a CEO does some things better and others slightly worse… then is it better or worse [on the whole]?” he rhetorically asked.

Yandex launched its AI system in May. The product is designed to create chatbots and AI assistants that generate, reorganize or sum up text-based information and solve other business-related issues. In September, the IT giant presented its upgraded version, YandexGPT 2, which the company claimed is able to solve more tasks and provide more precise answers to inquiries.

Developed by OpenAI and launched in 2022, ChatGPT has become the fastest-growing app in history, surpassing 100 million users in just two months. The product is a large language model-based chatbot that has a core function of mimicking a human conversationalist. It can also write and debug computer programs, compose anything from fairy tales to student essays, translate texts, and answer test questions.

The app’s success prompted other US technology giants, including Meta, Google and Microsoft, to start developing similar AI products of their own. The rapid development of AI technologies has also sparked fears of such systems getting out of control. In May, the US Congress grilled the CEO of OpenAI, Sam Altman, on the potential dangers posed by the technology.

Altman admitted that “significant harm” could be caused by the technology, which could “go quite wrong.” In particular, he said that AI’s powers of persuasion were one of his “greatest areas of concern.”
 
I was working on the subtitles for the Laura's interview with Elsa Schieder, and since the YT is very bad at making transcripts, I decided to use the help of ChatGPT for formatting the text. And it did a great job for the first batch of subtitles, but I had to use several batches because it cannot work with large amount of text. But with every next batch, he would do strange things that I would have to tell them specifically next time to not do it. At one point, he gave me more subtitle lines than what I gave him. I thought that he simply divided big lines into smaller ones, but now I just saw what he really did. He actually invented the whole section of the interview! And I never told him to do that.

Here is that part of the interview with Laura, imagined by ChatGPT:

I had a computer program that was scraping news from
All over the world

And uh, they were in this this this
Program was in a building and the building was located

Actually in between Nashville and Franklin, Tennessee
Uh, it was on the borders of the counties

And I would drive there and I would go to the
Data center and the guy would put a disk

In my hand, and I would
Drive back home and load it

So I started writing and
I said this guy's a psychopath

And I wrote an article about it
I said he's a psychopath

And then I listed, you know
All of the characteristics

And then I said this is why, you know, you've got a
Psychopathic CEO of the United States

I called it the, you know
The CEO of the United States

And I and I documented it
And I said, okay

This guy's a psychopath, and he's a murderer, and he's a liar
And he's this, that, and the other

And I wrote this article, and I put it
Up, and I said, you know, I've been researching this

And I've been studying psychopathy
And uh, it turns out that a lot of

People in high positions of power are psychopaths
And it's the thing to be in our world

You know if you're not
A psychopath, you're nobody

And that's what's wrong with our world, and that's why
We're going down the toilet

And so I put this article up, and it went uh
It was like, you know, viral before the term viral

Because it was getting millions of hits
A day on that one article

And people were sending it to everybody they
Knew, and it was just absolutely crazy

And then I got a phone call from a man
And he identified himself as a psychiatrist

And he said to me, I read your article
On psychopathy, and I want you to know

That I had diagnosed
George Bush as a psychopath

Years ago, and nobody wanted to
Listen to me, nobody would believe me

But I want you to know that
You're right, and I want to help you

And I said, well, that's really great, you know
Um, so tell me more about this

And he said, well, you
Know, I'm retired now

But I used to work for the FBI
And the CIA

And I was, you know,
Pretty high up

I mean, he was a a high-level
CIA psychiatrist, basically

And so he and I started
Corresponding, and he was sending me

Information, and he sent me this book
Political Ponerology

And um, he also invited me to attend
An international political ponerology conference

Which was being held in
Canada

And um, I said, well, that's
Really interesting

But I can't really afford
To go to Canada

And he said, well, you know,
We'll pay your way

And uh, that was like, wow, you know
So that was my introduction to ponerology

And so I went to this conference, and I met
A whole lot of really interesting people

And uh, I met a lot of
Uh, Canadians who were just

Crazy over the fact that an American was
Willing to come to Canada to learn about their work

Um, and I got to meet a lot of people
Who had been working on the problem for years

You know, they they didn't really call it
Ponerology per se, but they

Knew that something was
Really, really wrong in the world

And um, I started reading the book
Political Ponerology, and I was absolutely astounded

And the first thing that I that
I thought was amazing was that

They had actually documented that there are two types
Of psychopaths

There's your regular
Psychopath

And then there's what they call
A characteropathic psychopath

And a characteropathic psychopath is somebody
Who has been shaped

Who has been made into a psychopath
By the society

And by the conditions and the traumas
And the brutalities and so forth

They have been subjected to,
And so I'm reading this book

And I'm realizing that that's what we've
Got here is characteropathic psychopaths

Because I'm also
Reading about

The the genetics of
Psychopathy

And the genetics of psychopathy say that
It is a trait

And you cannot say that
Somebody's a psychopath

Or a characteropathic psychopath because
They have this trait

But unless it's actually
Triggered

By some circumstance, some experience
Some encounter, you know, something

That happens to them
They will not

Necessarily become a
Full-blown psychopath

And so this this book was just
Absolutely amazing

And I have it right here, and I'm
Just going to grab it

And I just want to show
Everybody the book

I should have had it ready, but

This is a this is a copy of
The book Political Ponerology

This is the second edition
Which is the improved edition

And this book was written by a Polish
Scientist by the name of Andrew Lobaczewski

And uh, it was really difficult
For him to write this book

And in fact, he wrote it
In Polish

And he was he was getting help
From his friend

And his friend was
Sending him files

Of the work as he wrote
It and they were

They were printing these files, and
They were sending them out

All over the world, and
So it actually got published

And it's just a really you know
Interesting book

And um, it it has it has
A lot of

Really wonderful material in it that explains the
Nature of evil and the nature of psychopathy

And how this this
Condition arises in in people

Uh, because basically you know
As I was saying

When you have a psychopath, you know, they are
Born with a certain genetic predisposition

But it's not triggered, you know
Unless there's a certain trauma or a certain

Experience in childhood
Or in you know

Or maybe even later in life
That sets them on this path

And this book explains how it works, and it's just
It's really an amazing amazing book

And I'm just I'm just really grateful to have
Had the opportunity to read it and to

To understand what's going on in the world
Because it's really it's really enlightening

And uh, so you know, I I I
Just went on from there

And um, and my husband and I we continued
To write about politics and about psychopathy

And we continued to write about all
Kinds of other things

Because once you start down this path of
Researching and understanding

You know, the nature of evil, it's
It's like a rabbit hole that you can't stop going down

And you you just you start seeing
All of these connections and all of these

Patterns and and you know, it's it's it's really
Fascinating, and it's also really

Terrifying because you realize that the
World is not what it appears to be

And that there are really dark forces at play
And that you know

It's it's it's it's a it's a struggle
To to to try to understand it

And to try to
You know, to try to

To try to make a difference, you know
To try to to try to

You know, to try to
Help people understand what's happening

Because there's there's so much
Deception and so much manipulation

And uh, and it's it's really you know,
It's it's it's it's a challenge

But um, but it's it's it's something that that
We feel is important to do, and so we continue

To to to write and to research
And to try to help people understand

You know, the nature of evil
And the nature of psychopathy

And and how it's affecting the world and
And what can be done about it, if anything

So that's that's that's kind of you know,
The story of how I got into all of this

And it's it's it's it's been
A really fascinating journey, and

And I feel really privileged to have
Had the opportunity to to to learn

And to to to to understand and to to to to
You know, try to help others understand as well.
 
Putin's remarks are interesting. Sure AI is important, but for countries. I'm glad he didn't serve the usual BS about it being good for humanity or something. It's a strategic asset for countries, like nuclear missiles. Are nuclear missiles good for humanity? Maybe maybe not, who knows.
On another subject, it appears that chatGPT, 3.5 and 4, are becoming less "accurate". One rumor is that OpenAI (which is not "Open") is degrading its free products to push people to buy the paying subscriptions, and the other is that they tried to censor it so much that they somehow broke it.
 
Putin's remarks are interesting. Sure AI is important, but for countries. I'm glad he didn't serve the usual BS about it being good for humanity or something. It's a strategic asset for countries, like nuclear missiles. Are nuclear missiles good for humanity? Maybe maybe not, who knows.
On another subject, it appears that chatGPT, 3.5 and 4, are becoming less "accurate". One rumor is that OpenAI (which is not "Open") is degrading its free products to push people to buy the paying subscriptions, and the other is that they tried to censor it so much that they somehow broke it.

Yeah, the primacy of AI is all based on code isn't it? And all code is prone to human error. I don't know exactly how it all works exactly but I've been thinking recently that AI isn't somehow immune to entropy.
 
Yeah, the primacy of AI is all based on code isn't it? And all code is prone to human error. I don't know exactly how it all works exactly but I've been thinking recently that AI isn't somehow immune to entropy.
More or less in the sens that the learned data play a very big part. The code part is essentially related to math. You look at the stored data and you compute probabilistic. The computer answer the question "which is the more probable output with this input when you take at example this particular set of data?"

This is not formal logic, this is more fuzzy logic so I think human errors in code would be "diluted".

What is sure is that:
1/ the set of data can be oriented and not neutral
2/ answers can be filtered

Interesting remark in any case.
 
Yeah, the primacy of AI is all based on code isn't it? And all code is prone to human error. I don't know exactly how it all works exactly but I've been thinking recently that AI isn't somehow immune to entropy.
The more complicated a system is, the more fragile. That's why modern cars with their increasing reliance on electronics and software freak me out.
 
On another subject, it appears that chatGPT, 3.5 and 4, are becoming less "accurate". One rumor is that OpenAI (which is not "Open") is degrading its free products to push people to buy the paying subscriptions, and the other is that they tried to censor it so much that they somehow broke it.
I am currently using the ChatGPT for some translation, and it is still very good if you know how to use it properly. You must give him text in small chunks, and you should take a little break between chunks. If you do the opposite, you can make him do crazy stuff, like I demonstrated in my previous post. So, I don't think that its code is broken, I think that it is intentionally programmed in such way for a free 3.5 version.
 
I guess it's filtered at the ISP (Internet Service Provider) level.
Have you tried changing the DNS resolver to Cloudflare? In Poland, RT is blocked on an ISP DNS level. Traffic is not blocked at all, if you can resolve the address.

How to set up Cloudflare on various devices:
 
Back
Top Bottom